Immune injection management system and method based on live pig health recognition

Abstract

The invention discloses an immune injection management system and method based on live pig health recognition, relates to the technical field of intelligent animal husbandry, and solves the technical problem that an existing live pig vaccine injection method is poor in injection effect due to the fact that the state of a live pig is not detected before vaccine injection. Comprising an ear tag identification module used for identifying an ear tag of a live pig to obtain ear tag information and querying live pig information; the data acquisition module is used for acquiring detection image data; the vaccine state recognition unit is used for generating a vaccine state according to the precautions information and the vaccine injection record information; the live pig state identification unit is used for identifying the live pig state when the vaccine state is qualified; generating a state score of the corresponding live pig according to the live pig information; when the state score is greater than the injectable state score, generating an acquisition signal; inputting the image data into an injection part identification model to obtain an injection part label; the injection module is used for injection when the label of the injection part is qualified; generating an injection allowing signal, and performing vaccine injection on the live pigs; and the vaccine injection effect is effectively improved.

Description

Technical Field

[0001] The present application belongs to the field of intelligent animal husbandry technology, and relates to intelligent animal husbandry technology, specifically a management system and method based on pig health identification and immunization injection. Background Art

[0002] Animal husbandry is an important industry related to the national economy and people's livelihood, and the prevention and control of animal diseases is a key task. With the increasingly widespread application of artificial intelligence in animal husbandry, higher requirements are placed on the prevention and control of animal epidemics.

[0003] The prior art (the announcement number is CN113599013B invention patent) discloses a pig vaccine injection robot and a vaccine injection method, which solves the technical problems of the existing manual injection method of pig vaccination, which is labor-intensive, high labor cost, low injection efficiency, and easy to miss or inject by mistake. It includes a wheeled mobile platform, a six-degree-of-freedom collaborative robotic arm, a needle-free syringe, an RGB-D camera, a camera bracket and a control system. The six-degree-of-freedom collaborative robotic arm is connected to the wheeled mobile platform, the needle-free syringe is provided with a handle and an injection head, and the handle is fixedly connected to the end flange of the six-degree-of-freedom collaborative robotic arm; the camera bracket is connected to the end of the six-degree-of-freedom collaborative robotic arm, and the RGB-D camera is fixedly connected to the camera bracket; the wheeled mobile platform is provided with a mobile platform main controller and a motor driver, and the motor driver is connected to the mobile platform main controller.

[0004] The above-mentioned method for swine vaccination, by numbering each pig, records the pigs that have been vaccinated to prevent missed vaccinations and wrong vaccinations. The health status of the pigs during vaccination and the cleanliness of the injection site will affect the effect of the vaccination. If the vaccine is injected when the health status of the pigs is poor, it may cause severe death of the pigs. If the vaccine is injected when there are too many stains at the injection site, due to the use of needle-free injection, part of the vaccine will be blocked by the stains, thereby affecting the injection effect of the vaccine. Therefore, a management system and method for swine health identification and immunization injection is needed. Summary of the invention

[0005] The present application aims to solve at least one of the technical problems existing in the prior art; to this end, the present application proposes a pig health identification immunization injection management system and method, which is used to solve the technical problem that the existing pig vaccine injection method does not detect the status of the pigs before injecting the vaccine, resulting in poor injection effect.

[0006] To achieve the above-mentioned purpose, the first aspect of the present application provides a pig health identification and immunization injection management system, including: an ear tag identification module, a data acquisition module, a state identification module, an injection module, an injection inspection module, a cloud service platform, a database and a display module;

[0007] The ear tag recognition module is used to identify the ear tag of the pig to obtain the ear tag information, and upload the ear tag information to the cloud service platform to query and obtain the pig information of the corresponding pig; the pig information includes vaccination record information, historical normal physical sign data, and real-time physical sign data, and the physical sign data includes body temperature data, heart rate data, respiratory rate data and ambient temperature;

[0008] The data acquisition module is used to acquire acquisition signals and collect detection image data of the injection site of the live pig;

[0009] The state recognition module includes a vaccine state recognition unit and a pig state recognition unit;

[0010] The vaccine status identification unit: obtains the precaution information of the vaccine to be injected; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information;

[0011] The pig status recognition unit: when the vaccine status is qualified; generates a status score of the corresponding pig according to the pig information; when the status score is greater than the injectable status score in the precaution information, generates a collection signal; and obtains image data, inputs the image data into the injection part recognition model to obtain an injection part label; the injection part recognition model is obtained by training the artificial intelligence model;

[0012] The injection module: obtains the injection part label, and when the injection part label is qualified; generates an injection permission signal, and performs vaccine injection on the live pig.

[0013] The present application obtains the precaution information of the vaccine to be injected and the ear tag information of the target pig, and queries the corresponding pig information according to the ear tag information; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information; when the vaccine status is qualified; generates the corresponding pig status score according to the pig information; when the status score is greater than the injectable status score in the precaution information, collects the detection image data of the injection site of the pig; inputs the image data into the injection site recognition model to obtain the injection site label; when the injection site label is qualified; generates an injection permission signal to vaccinate the pig; before the pig is vaccinated, detects the pig's physical condition and the cleanliness of the injection site, and injects the pig if it meets the injection standards; at the same time, uses a needle-free injection method to avoid cross infection, which effectively improves the effect of vaccine injection.

[0014] Preferably, generating the vaccine status according to the precaution information and the vaccine injection record information includes:

[0015] Extract several conflicting vaccine categories and their corresponding conflicting times in the precautions, as well as several injected vaccine categories and their corresponding interval times in the vaccine injection record information;

[0016] Determine whether there is a vaccine category that is the same as the conflicting vaccine category among the categories of vaccines that have been administered;

[0017] If yes, then when the interval time corresponding to the injected vaccine category is less than the conflict time, the vaccine status is set to unqualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information; otherwise, the vaccine status is set to qualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information;

[0018] If not, the vaccine status will be set to qualified; and all conflicting vaccine categories, conflicting times, and injected vaccine categories and interval times will be packaged into vaccine supporting information.

[0019] Preferably, generating a status score of a corresponding pig according to the pig information includes:

[0020] Extracting real-time vital sign data and historical normal vital sign data from the live pig information; the historical normal vital sign data and the vital sign data have the same data type, both including body temperature data, heart rate data, respiratory rate data and ambient temperature;

[0021] The historical normal vital sign data with the same ambient temperature are divided into a vital sign characteristic group; the average value of the body temperature data in each vital sign characteristic group is used as the normal reference body temperature of the corresponding vital sign data group; the average value of the heart rate data in each vital sign characteristic group is used as the normal reference heart rate of the corresponding vital sign characteristic group; the average value of the respiratory rate data in each vital sign characteristic group is used as the normal reference respiratory rate of the corresponding vital sign characteristic group; the normal reference body temperature, the normal reference heart rate and the normal reference respiratory rate are integrated into the reference vital sign data of the corresponding vital sign characteristic group; the reference vital sign data of each vital sign characteristic group of the corresponding live pigs are obtained in sequence;

[0022] Generate a pig condition score based on reference vital sign data and real-time vital sign data.

[0023] Preferably, generating a pig status score based on reference vital sign data and real-time vital sign data comprises:

[0024] Extract the ambient temperature from the physical sign data, mark the body temperature data as ST, mark the heart rate data as SX, and mark the respiratory rate data as SH; match the corresponding physical sign feature group according to the ambient temperature, obtain the reference physical sign data corresponding to the physical sign feature group, and mark the normal reference body temperature in the reference physical sign data as ZT, the normal reference heart rate as ZX, and the normal reference respiratory rate as ZH;

[0025] The current status score ZP of the corresponding pig is calculated by the formula ZP=α1×exp(-|ST-ZT| / ZT)+α2×exp(-|SX-ZX| / ZX)+α3×exp(-|SH-ZH| / ZH); wherein α1, α2 and α3 are weight coefficients, all of which are set based on experience.

[0026] Preferably, the injection part recognition model is obtained by training an artificial intelligence model, including:

[0027] Acquire a number of historical detection image data and their corresponding injection part labels through a database; the injection part labels include qualified and unqualified; integrate the historical detection image data and their corresponding injection part labels into a number of training data and inspection data;

[0028] The artificial intelligence model is trained using training data; the trained artificial intelligence model is tested using test data, and finally an injection part recognition model is obtained, whose input is the detection image data and whose output is the corresponding injection part label; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

[0029] Preferably, the injection module also generates a test acquisition signal after the injection is completed;

[0030] The data acquisition module is also used to obtain the inspection acquisition signal and collect the inspection image data of the injection site of the live pig;

[0031] The injection inspection module: acquires inspection image data, inputs the inspection image data into an abnormality recognition model to obtain an inspection label and its corresponding abnormal area; when the inspection label is an injection abnormality, generates a re-injection signal according to the abnormal area, and re-injects the vaccine into the pig according to the injection pressure; when the inspection label is an allergy abnormality, generates an allergy alarm signal; and sends the allergy alarm signal and the inspection image data to the display module; the abnormality recognition model is obtained through training of an artificial intelligence model.

[0032] Preferably, the anomaly recognition model is obtained by training an artificial intelligence model, including:

[0033] Obtain a number of historical inspection image data, as well as the corresponding inspection labels and abnormal areas through a database; the inspection labels include injection abnormality, allergic abnormality and normal; the abnormal area is the area where the abnormality appears in the inspection image, and when the inspection label is normal, the abnormal area is zero; when the inspection label is injection abnormality, the abnormal area is the area where the vaccine leaks on the pig's skin; when the inspection label is allergic abnormality, the abnormal area is the area where the pig is allergic at the injection site; integrate a number of historical inspection image data, inspection labels and abnormal areas into a number of training data and inspection data;

[0034] The artificial intelligence model is trained using the training data; the trained artificial intelligence model is tested using the test data, and finally an abnormality recognition model is obtained whose input is the test image data and whose output is the corresponding test label and abnormal area; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

[0035] The present application detects the injection part corresponding to the injected pig to determine whether the vaccine is not completely injected into the pig's body due to insufficient injection pressure; and generates a corresponding alarm signal to conduct timely detection of the injection result, so that abnormalities in the injection result can be discovered in time.

[0036] Preferably, generating a re-injection signal according to the abnormal area includes:

[0037] Obtain the original injection pressure and mark it as YY; obtain the abnormal area and mark it as XM; determine whether the abnormal area is greater than the abnormal area threshold;

[0038] Yes, the adjusted injection pressure TY is calculated by the formula TY=YY+exp(k×XM×FG / TJ-1)×KY; the adjusted injection pressure replaces the original injection pressure to generate a re-injection signal; wherein FG is the vaccine adhesion thickness; k is the adjustment coefficient, TJ is the volume of the vaccine to be injected, and KY is the adjustable injection pressure;

[0039] If not, an injection leakage alarm signal is generated, and the injection leakage alarm signal and the inspection image data are sent to the display module.

[0040] Another aspect of the present application provides a method for managing immunization injection based on pig health identification, comprising the following steps:

[0041] Step 1: Obtain the precautions information of the vaccine to be injected and the ear tag information of the target pig, and query the corresponding pig information according to the ear tag information, wherein the pig information includes the vaccine injection record information and the historical normal physical sign data, as well as the real-time physical sign data;

[0042] Step 2: Extract the vaccination record information in the pig information; generate the vaccination status according to the precaution information and the vaccination record information; determine whether the vaccination status is qualified; if yes, proceed to step 3; if no, generate a vaccination conflict signal and proceed to step 6;

[0043] Step 3: Generate a status score for the corresponding live pig based on the live pig information; determine whether the status score is greater than the injectable status score in the precaution information; if yes, collect the detection image data of the live pig's injection site and proceed to step 4; if no, generate a live pig status abnormality signal and proceed to step 6;

[0044] Step 4: input the image data into the injection part recognition model to obtain the injection part label; the injection part recognition model is obtained by training the artificial intelligence model; determine whether the image label is qualified; if yes, proceed to step 5; if no, generate an injection part abnormal signal and proceed to step 6;

[0045] Step 5: Generate an injection permission signal and vaccinate the pigs;

[0046] Step 6: Send the vaccine injection conflict signal, or the pig status abnormal signal, or the injection part abnormal signal to the display module; vaccination of pigs is not allowed.

[0047] Preferably, the step five further comprises: after the injection is completed, collecting inspection image data of the injection site of the live pig; inputting the inspection image data into an abnormality recognition model to obtain an inspection label and its corresponding abnormal area; the abnormality recognition model is obtained by training an artificial intelligence model;

[0048] When the inspection tag indicates an injection abnormality, a re-injection signal is generated based on the abnormal area, and the pigs are re-vaccinated based on the injection pressure;

[0049] When the inspection tag is allergic abnormal, an allergy alarm signal is generated; and the allergy alarm signal and the inspection image data are sent to the display module.

[0050] Compared with the prior art, the beneficial effects of this application are:

[0051] 1. The present application obtains the precaution information of the vaccine to be injected and the ear tag information of the target pig, and queries the corresponding pig information according to the ear tag information; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information; when the vaccine status is qualified; generates the corresponding pig status score according to the pig information; when the status score is greater than the injectable status score in the precaution information, collects the detection image data of the injection site of the pig; inputs the image data into the injection site recognition model to obtain the injection site label; when the injection site label is qualified; generates an injection permission signal to vaccinate the pig; before the pig is vaccinated, detects the pig's physical condition and the cleanliness of the injection site, and injects the pig if it meets the injection standards; at the same time, uses a needle-free injection method to avoid cross infection, which effectively improves the effect of vaccine injection.

[0052] 2. The present application detects the injection part corresponding to the injected pig to determine whether the vaccine is not completely injected into the pig's body due to insufficient injection pressure; and generates a corresponding alarm signal to conduct timely detection of the injection result, so that abnormalities in the injection result can be discovered in time. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 This is a schematic diagram of the principle of the pig health identification and immunization injection management system in this application;

[0055] Figure 2 This is a method step diagram for the pig health identification immunization injection management method in this application. DETAILED DESCRIPTION

[0056] The technical solution of the present application will be described clearly and completely below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0057] See also Figure 1, the first embodiment of the present application provides a system for immunization injection management based on pig health identification, including: an ear tag identification module, a data acquisition module, a state identification module, an injection module, an injection inspection module, a cloud service platform, a database and a display module;

[0058] Ear tag recognition module: used to identify the ear tag of the pig to obtain the ear tag information, which includes the unique number of the corresponding pig; and upload the ear tag information to the cloud service platform to query and obtain the pig information of the corresponding pig; the ear tag of the pig contains the unique number corresponding to the pig; the cloud service platform records the information of each pig corresponding to the ear tag number, as well as the real-time collection of the pig's body temperature data, heart rate data and respiratory rate data through the smart ear tag, and the real-time acquisition of the ambient temperature of the pig house through the temperature collection equipment; the pig information includes vaccination record information, historical normal vital signs data, and real-time vital signs data, which includes body temperature data, heart rate data, respiratory rate data and ambient temperature; the historical normal vital signs data is the same as the data type of the vital signs data, indicating the body temperature data, heart rate data and respiratory rate data collected when the pig is in a normal state, as well as the corresponding environmental data.

[0059] Data acquisition module: obtains the acquisition signal and collects the detection image data of the injection site of the live pig through the data acquisition device connected to it; the data acquisition device is a high-definition camera, etc.;

[0060] The status recognition module includes a vaccine status recognition unit and a pig status recognition unit;

[0061] Vaccine status identification unit: obtains precaution information for vaccines to be injected; precaution information includes several conflicting vaccine categories and their corresponding conflicting times, as well as injectable status scores. Conflicting vaccine categories are vaccine categories that cannot be injected with vaccines to be injected within a certain period of time, and conflicting time is the time interval between two injections that will cause conflicting vaccines to be injected safely; injectable status scores are standard scores for the health status of pigs according to the inspection settings; extracts vaccine injection record information from pig information; vaccine injection record information is the vaccines that have been injected into the pigs and the time interval between injections and the current time; generates vaccine status based on precaution information and vaccine injection record information;

[0062] Pig status identification unit: when the vaccine status is qualified; generate the corresponding pig status score according to the pig information; the status score is the score of the current health status of the pig; when the status score is greater than the injectable status score in the precautions information, it means that the health status of the pig meets the conditions for vaccine injection; then generate an acquisition signal; and obtain image data, input the image data into the injection part identification model to obtain the injection part label; the injection part identification model is obtained by training the artificial intelligence model;

[0063] The injection module: obtains the injection part label, and when the injection part label is qualified; generates an injection permission signal, and performs vaccination on the pig; and uploads the injected vaccine and the injection time to the cloud service platform. For the specific injection process, please refer to the patent for invention number CN113599013B, the invention name of which is a method for vaccination using a pig vaccination robot.

[0064] This embodiment obtains the precaution information of the vaccine to be injected and the ear tag information of the target pig, and queries the corresponding pig information according to the ear tag information; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information; when the vaccine status is qualified; generates the corresponding pig status score according to the pig information; when the status score is greater than the injectable status score in the precaution information, collects the detection image data of the injection site of the pig; inputs the image data into the injection site recognition model to obtain the injection site label; when the injection site label is qualified; generates an injection permission signal to vaccinate the pig; before the pig is vaccinated, detects the pig's physical condition and the cleanliness of the injection site, and injects the pig if it meets the injection standards; at the same time, uses a needle-free injection method to avoid cross infection, which effectively improves the effect of vaccine injection.

[0065] Generate vaccine status according to precaution information and vaccine injection record information, including: extracting several conflicting vaccine categories and their corresponding conflicting times in precaution information, and several injected vaccine categories and their corresponding interval times in vaccine injection record information;

[0066] Determine whether there is a vaccine category that is the same as the conflicting vaccine category among the categories of vaccines that have been administered;

[0067] If yes, then when the interval time corresponding to the injected vaccine category is less than the conflict time, the vaccine status is set to unqualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information; otherwise, the vaccine status is set to qualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information; for example, when injecting swine fever vaccine, the conflict vaccine category can be swine fever vaccine, and its corresponding conflict time is one year; if the interval between the last injection of swine fever vaccine and this injection is one month, the vaccine status needs to be set to unqualified this time;

[0068] If not, the vaccine status will be set to qualified; all conflicting vaccine categories, conflicting times, and injected vaccine categories and interval times will be packaged into vaccine supporting information; and the generated vaccine supporting information will be sent to the display module for reference by relevant personnel.

[0069] This embodiment detects whether there are conflicting vaccines, thereby avoiding repeated injections of the same vaccine and the possibility of illegal injections of multiple vaccines that cannot be injected within a certain period of time.

[0070] Generate a status score of the corresponding pig according to the pig information, including: extracting real-time vital sign data and historical normal vital sign data from the pig information; the data types of the historical normal vital sign data and the vital sign data are the same, both including body temperature data, heart rate data, respiratory rate data and ambient temperature; the body temperature data is the current body temperature of the pig collected by the smart ear tag, the heart rate data is the current heart rate of the pig collected by the smart ear tag; the respiratory rate data is the current respiratory rate of the pig collected by the smart ear tag; or the pig vital sign data is measured in real time by existing equipment, such as a thermometer, a heart rate detector, a respiratory rate detector, etc., and then uploaded to the cloud service platform and associated with the unique identification number of the pig;

[0071] The historical normal vital sign data with the same ambient temperature are divided into one vital sign feature group; it can be understood that the same ambient temperature in this embodiment means that the integer part of the ambient temperature is the same; in another embodiment, the vital sign data group is adjusted according to the historical normal vital sign data in each vital sign feature group; specifically, when the difference between the maximum value and the minimum value of any one of the body temperature data, heart rate data and respiratory rate data in the vital sign data group is greater than the preset difference threshold, the corresponding vital sign data group is split, specifically, the ambient temperature range corresponding to the vital sign data group is equally divided into two parts, and the historical normal vital sign data within the two ambient temperature ranges are divided into one vital sign feature group; continue to judge whether the difference between the maximum value and the minimum value of any one of the body temperature data, heart rate data and respiratory rate data in the vital sign data group is greater than the corresponding preset difference threshold; if yes, continue to equally divide the ambient temperature range corresponding to the corresponding vital sign data group into two parts, otherwise, end the adjustment operation.

[0072] The average value of the body temperature data in each physical sign characteristic group is used as the normal reference body temperature of the corresponding physical sign data group; the average value of the heart rate data in each physical sign characteristic group is used as the normal reference heart rate of the corresponding physical sign characteristic group; the average value of the respiratory rate data in each physical sign characteristic group is used as the normal reference respiratory rate of the corresponding physical sign characteristic group; the normal reference body temperature, the normal reference heart rate and the normal reference respiratory rate are integrated into the reference physical sign data of the corresponding physical sign characteristic group; the reference physical sign data of each physical sign characteristic group of the corresponding live pig is obtained in sequence;

[0073] Generate a pig condition score based on reference vital sign data and real-time vital sign data.

[0074] In this embodiment, the reference vital sign data of the corresponding pigs at various ambient temperatures are generated through the historical normal vital sign data of the pigs, and the historical data is referred to so that when the corresponding status score of the pigs is subsequently generated, the score is more accurate.

[0075] Generate a status score of the live pig according to the reference vital sign data and the real-time vital sign data, including: extracting the ambient temperature in the vital sign data, marking the body temperature data as ST, marking the heart rate data as SX, and marking the respiratory rate data as SH; matching the corresponding vital sign feature group according to the ambient temperature, obtaining the reference vital sign data corresponding to the vital sign feature group, and marking the normal reference body temperature in the reference vital sign data as ZT, the normal reference heart rate as ZX, and the normal reference respiratory rate as ZH;

[0076] The current status score ZP of the corresponding pig is calculated by the formula ZP=α1×exp(-|ST-ZT| / ZT)+α2×exp(-|SX-ZX| / ZX)+α3×exp(-|SH-ZH| / ZH); wherein α1, α2 and α3 are weight coefficients, all of which are set based on experience.

[0077] This embodiment generates a corresponding status score based on the gap between the current vital sign data of the pig and the reference vital sign data under the same environment. When the gap between the reference vital sign data and the current vital sign data is larger, it means that the deviation between the current status of the pig and the healthy status is larger, and the corresponding status score is smaller; when the gap between the reference vital sign data and the current vital sign data is smaller, it means that the current status of the pig is closer to the healthy status, and the corresponding status score is larger; according to the status score of the pig, it is convenient to judge whether the current status of the pig is abnormal.

[0078] The injection part recognition model is obtained through artificial intelligence model training, including:

[0079] Acquire a number of historical detection image data and their corresponding injection part labels through a database; the injection part labels include qualified and unqualified; integrate the historical detection image data and their corresponding injection part labels into a number of training data and inspection data;

[0080] The artificial intelligence model is trained using training data; the trained artificial intelligence model is tested using test data, specifically, historical image data in the test data is input into the trained artificial intelligence model to obtain an injection part label as output; it is determined whether the injection part label is consistent with the injection part label recorded in the test data, if so, subsequent test data testing is performed, if not, relevant parameters are adjusted and the test of this group of test data is continued; until the test results of several test data meet the relevant standards, an injection part recognition model is finally obtained with the input being the detection image data and the output being the corresponding injection part label; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

[0081] After the injection is completed, the injection module will also generate a test acquisition signal;

[0082] The data acquisition module is also used to obtain the inspection acquisition signal and collect the inspection image data of the injection site of the live pig;

[0083] Injection inspection module: acquire inspection image data, input the inspection image data into the abnormality recognition model to obtain the inspection label and its corresponding abnormal area; when the inspection label is an injection abnormality, generate a re-injection signal according to the abnormal area, and re-inject the vaccine into the pig according to the injection pressure; when the inspection label is an allergy abnormality, generate an allergy alarm signal; and send the allergy alarm signal and the inspection image data to the display module; the abnormality recognition model is obtained through artificial intelligence model training.

[0084] In this embodiment, the injection part corresponding to the injected pig is detected to determine whether the vaccine is not completely injected into the pig's body due to insufficient injection pressure; and a corresponding alarm signal is generated to perform timely detection of the injection result, so that abnormalities in the injection result can be discovered in time, thereby improving the efficiency of the injection.

[0085] The anomaly recognition model is obtained through artificial intelligence model training, including:

[0086] Obtain a number of historical inspection image data, as well as the corresponding inspection labels and abnormal areas through a database; the inspection labels include injection abnormality, allergic abnormality and normal; the abnormal area is the area where the abnormality appears in the inspection image, and when the inspection label is normal, the abnormal area is zero; when the inspection label is injection abnormality, the abnormal area is the area where the vaccine leaks on the pig's skin; when the inspection label is allergic abnormality, the abnormal area is the area where the pig is allergic at the injection site; integrate a number of historical inspection image data, inspection labels and abnormal areas into a number of training data and inspection data;

[0087] The artificial intelligence model is trained using the training data; the trained artificial intelligence model is tested using the test data, and finally an abnormality recognition model is obtained whose input is the test image data and whose output is the corresponding test label and abnormal area; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

[0088] Generating a re-injection signal according to the abnormal area includes: obtaining the original injection pressure and marking it as YY; obtaining the abnormal area and marking it as XM; judging whether the abnormal area is greater than the abnormal area threshold;

[0089] Yes, the adjusted injection pressure TY is calculated through the formula TY=YY+exp(k×XM×FG / TJ-1)×KY; the original injection pressure is replaced by the adjusted injection pressure. The more vaccine is exposed, the worse the injection effect is. Properly increasing the injection pressure can speed up the injection speed and injection intensity, so as to facilitate the complete injection of the vaccine into the pig and generate a re-injection signal; wherein FG is the vaccine attachment thickness; k is the adjustment coefficient, TJ is the volume of vaccine to be injected, and KY is the adjustable injection pressure; and the adjusted injection pressure is transmitted to the cloud service platform for matching with the corresponding pig information. When the pig is injected next time, the adjusted injection pressure is obtained and injection is performed according to the adjusted injection pressure.

[0090] If not, an injection leakage alarm signal is generated, and the injection leakage alarm signal and the inspection image data are sent to the display module.

[0091] This embodiment adjusts different injection pressures according to the vaccine leakage at the injection site to achieve a higher injection effect.

[0092] See also Figure 2 Another aspect of the present application provides a method for managing immunization injection based on pig health identification, comprising the following steps:

[0093] Step 1: Obtain the precautions information of the vaccine to be injected and the ear tag information of the target pig, and query the corresponding pig information according to the ear tag information, wherein the pig information includes the vaccine injection record information and the historical normal physical sign data, as well as the real-time physical sign data;

[0094] Step 2: Extract the vaccination record information in the pig information; generate the vaccination status according to the precaution information and the vaccination record information; determine whether the vaccination status is qualified; if yes, proceed to step 3; if no, generate a vaccination conflict signal and proceed to step 6;

[0095] Step 3: Generate a status score for the corresponding live pig based on the live pig information; determine whether the status score is greater than the injectable status score in the precaution information; if yes, collect the detection image data of the live pig's injection site and proceed to step 4; if no, generate a live pig status abnormality signal and proceed to step 6;

[0096] Step 4: input the image data into the injection part recognition model to obtain the injection part label; the injection part recognition model is obtained by training the artificial intelligence model; determine whether the image label is qualified; if yes, proceed to step 5; if no, generate an injection part abnormal signal and proceed to step 6;

[0097] Step 5: Generate an injection permission signal and vaccinate the pigs;

[0098] Step 6: Send the vaccine injection conflict signal, or the pig status abnormal signal, or the injection part abnormal signal to the display module; vaccination of pigs is not allowed.

[0099] Step five also includes: after the injection is completed, collecting inspection image data of the injection site of the live pig; inputting the inspection image data into an abnormality recognition model to obtain an inspection label and its corresponding abnormal area; the abnormality recognition model is obtained by training an artificial intelligence model;

[0100] When the inspection tag indicates an injection abnormality, a re-injection signal is generated based on the abnormal area, and the pigs are re-vaccinated based on the injection pressure;

[0101] When the inspection tag is allergic abnormal, an allergy alarm signal is generated; and the allergy alarm signal and the inspection image data are sent to the display module.

[0102] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0103] How this application works:

[0104] This embodiment obtains the precaution information of the vaccine to be injected and the ear tag information of the target pig, and queries the corresponding pig information according to the ear tag information; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information; when the vaccine status is qualified; generates the corresponding pig status score according to the pig information; when the status score is greater than the injectable status score in the precaution information, collects the detection image data of the injection site of the pig; inputs the image data into the injection site recognition model to obtain the injection site label; when the injection site label is qualified; generates an injection permission signal to vaccinate the pig; before the pig is vaccinated, detects the pig's physical condition and the cleanliness of the injection site, and injects the pig if it meets the injection standards; at the same time, uses a needle-free injection method to avoid cross infection, which effectively improves the effect of vaccine injection.

[0105] The above embodiments are only used to illustrate the technical method of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical method of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present application.

Claims

1. A pig health identification and immunization injection management system, comprising: Ear tag recognition module, data acquisition module, state recognition module, injection module, injection inspection module, cloud service platform, database and display module; characterized in that, The ear tag recognition module is used to identify the ear tag of the pig to obtain the ear tag information, and upload the ear tag information to the cloud service platform to query and obtain the pig information of the corresponding pig; The data acquisition module is used to acquire acquisition signals and collect detection image data of the injection site of the live pig; The state recognition module includes a vaccine state recognition unit and a pig state recognition unit; The vaccine status identification unit: obtains the precaution information of the vaccine to be injected; extracts the vaccine injection record information in the pig information; generates the vaccine status according to the precaution information and the vaccine injection record information; The pig status recognition unit: when the vaccine status is qualified; generates a status score of the corresponding pig according to the pig information; when the status score is greater than the injectable status score in the precaution information, generates a collection signal; and obtains image data, inputs the image data into the injection part recognition model to obtain an injection part label; the injection part recognition model is obtained by training the artificial intelligence model; The injection module: obtains the injection part label, and when the injection part label is qualified; generates an injection permission signal, and performs vaccine injection on the live pig.

2. The system for managing pig health identification and immunization injection according to claim 1 is characterized in that: The generating of the vaccine status according to the precaution information and the vaccine injection record information includes: Extract several conflicting vaccine categories and their corresponding conflicting times in the precautions, as well as several injected vaccine categories and their corresponding interval times in the vaccine injection record information; Determine whether there is a vaccine category that is the same as the conflicting vaccine category among the vaccines that have been administered; If yes, then when the interval time corresponding to the injected vaccine category is less than the conflict time, the vaccine status is set to unqualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information; otherwise, the vaccine status is set to qualified; and the corresponding conflict vaccine category, conflict time, and injected vaccine category and interval time are packaged into vaccine supporting information; If not, the vaccine status will be set to qualified; and all conflicting vaccine categories, conflicting times, and injected vaccine categories and interval times will be packaged into vaccine supporting information.

3. The system for managing pig health identification and immunization injection according to claim 1 is characterized in that: Generating a status score of a corresponding pig according to the pig information includes: Extracting real-time vital sign data and historical normal vital sign data from the live pig information; the historical normal vital sign data and the vital sign data have the same data type, both including body temperature data, heart rate data, respiratory rate data and ambient temperature; The historical normal vital sign data with the same ambient temperature are divided into a vital sign characteristic group; the average value of the body temperature data in each vital sign characteristic group is used as the normal reference body temperature of the corresponding vital sign data group; the average value of the heart rate data in each vital sign characteristic group is used as the normal reference heart rate of the corresponding vital sign characteristic group; the average value of the respiratory rate data in each vital sign characteristic group is used as the normal reference respiratory rate of the corresponding vital sign characteristic group; the normal reference body temperature, the normal reference heart rate and the normal reference respiratory rate are integrated into the reference vital sign data of the corresponding vital sign characteristic group; the reference vital sign data of each vital sign characteristic group of the corresponding live pigs are obtained in sequence; Generate a pig condition score based on reference vital sign data and real-time vital sign data.

4. A system for managing pig health identification and immune injection according to claim 3, characterized in that: The method of generating a pig's status score based on the reference vital sign data and the real-time vital sign data includes: Extract the ambient temperature from the physical sign data, mark the body temperature data as ST, mark the heart rate data as SX, and mark the respiratory rate data as SH; match the corresponding physical sign feature group according to the ambient temperature, obtain the reference physical sign data corresponding to the physical sign feature group, and mark the normal reference body temperature in the reference physical sign data as ZT, the normal reference heart rate as ZX, and the normal reference respiratory rate as ZH; The current status score ZP of the corresponding pig is calculated by the formula ZP=α1×exp(-|ST-ZT| / ZT)+α2×exp(-|SX-ZX| / ZX)+α3×exp(-|SH-ZH| / ZH); wherein α1, α2 and α3 are weight coefficients, all of which are set based on experience.

5. The system for immunization injection based on pig health identification according to claim 1 is characterized in that: The injection part recognition model is obtained by training an artificial intelligence model, and includes: Acquire a number of historical detection image data and their corresponding injection part labels through a database; the injection part labels include qualified and unqualified; integrate the historical detection image data and their corresponding injection part labels into a number of training data and inspection data; The artificial intelligence model is trained using training data; the trained artificial intelligence model is tested using test data, and finally an injection part recognition model is obtained, whose input is the detection image data and whose output is the corresponding injection part label; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

6. The system for immunization injection based on pig health identification according to claim 1 is characterized in that: After the injection is completed, the injection module will also generate a test acquisition signal; The data acquisition module is also used to obtain the inspection acquisition signal and collect the inspection image data of the injection site of the live pig; The injection inspection module: acquires inspection image data, and inputs the inspection image data into an abnormality recognition model to obtain an inspection label and its corresponding abnormal area; When the inspection tag is an injection abnormality, a re-injection signal is generated according to the abnormal area, and the pigs are re-vaccinated according to the injection pressure; when the inspection tag is an allergy abnormality, an allergy alarm signal is generated; and the allergy alarm signal and the inspection image data are sent to the display module; the abnormality recognition model is obtained through artificial intelligence model training.

7. The system for managing pig health identification and immune injection according to claim 6 is characterized in that: The anomaly recognition model is obtained through artificial intelligence model training, including: Acquire a number of historical inspection image data, and their corresponding inspection labels and abnormal areas through a database; the inspection labels include injection abnormality, allergy abnormality and normal; integrate the number of historical inspection image data, inspection labels and abnormal areas into a number of training data and inspection data; The artificial intelligence model is trained using the training data; the trained artificial intelligence model is tested using the test data, and finally an abnormality recognition model is obtained whose input is the test image data and whose output is the corresponding test label and abnormal area; wherein the artificial intelligence model includes a BP neural network model and an RBF neural network model.

8. The system for managing pig health identification and immune injection according to claim 6 is characterized in that: The step of generating a reinjection signal according to the abnormal area comprises: Obtain the original injection pressure and mark it as YY; obtain the abnormal area and mark it as XM; determine whether the abnormal area is greater than the abnormal area threshold; Yes, the adjusted injection pressure TY is calculated by the formula TY=YY+exp(k×XM×FG / TJ-1)×KY; the adjusted injection pressure replaces the original injection pressure to generate a re-injection signal; wherein FG is the vaccine adhesion thickness; k is the adjustment coefficient, TJ is the volume of the vaccine to be injected, and KY is the adjustable injection pressure; If not, an injection leakage alarm signal is generated, and the injection leakage alarm signal and the inspection image data are sent to the display module.

9. A method for managing immunization injection based on healthy identification of live pigs, applied to the operation of a management system based on healthy identification of live pigs as claimed in any one of claims 1 to 8; characterized in that: The following steps are involved: Step 1: Obtain the precautions information of the vaccine to be injected and the ear tag information of the target pig, and query the corresponding pig information according to the ear tag information, wherein the pig information includes the vaccine injection record information and the historical normal physical sign data, as well as the real-time physical sign data; Step 2: Extract the vaccination record information in the pig information; generate the vaccination status according to the precaution information and the vaccination record information; determine whether the vaccination status is qualified; if yes, proceed to step 3; if no, generate a vaccination conflict signal and proceed to step 6; Step 3: Generate a status score for the corresponding live pig based on the live pig information; determine whether the status score is greater than the injectable status score in the precaution information; if yes, collect the detection image data of the live pig's injection site and proceed to step 4; if no, generate a live pig status abnormality signal and proceed to step 6; Step 4: input the image data into the injection part recognition model to obtain the injection part label; the injection part recognition model is obtained by training the artificial intelligence model; determine whether the image label is qualified; if yes, proceed to step 5; if no, generate an injection part abnormal signal and proceed to step 6; Step 5: Generate an injection permission signal and vaccinate the pigs; Step 6: Send the vaccine injection conflict signal, or the pig status abnormal signal, or the injection part abnormal signal to the display module; vaccination of pigs is not allowed.

10. A method for managing pig health identification and immunization injection according to claim 9, characterized in that: The step five also includes: after the injection is completed, collecting inspection image data of the injection site of the live pig; inputting the inspection image data into an abnormality recognition model to obtain an inspection label and its corresponding abnormal area; the abnormality recognition model is obtained by training an artificial intelligence model; When the inspection tag indicates an injection abnormality, a re-injection signal is generated based on the abnormal area, and the pigs are re-vaccinated based on the injection pressure; When the inspection tag is allergic abnormal, an allergy alarm signal is generated; and the allergy alarm signal and the inspection image data are sent to the display module.

Images